Fast surgery for quantum LDPC codes

aqora / Fast surgery for quantum LDPC codes

Public

About dataset version

This dataset accompanies the paper Fast Surgery for Quantum LDPC Codes (2025) and consolidates the simulation results from all three CSV files provided by the authors on Zenodo (record 17220908). It provides empirical data used to benchmark the performance of lattice-surgery-based fault-tolerant operations for quantum LDPC codes under different noise configurations.

Dataset contents

Column	Type	Description
`shots`	int64	Number of Monte Carlo simulation shots performed
`failures`	int64	Number of failed logical operations during simulation
`task_id`	string	Unique identifier for the simulation task
`core_seconds`	float64	Total compute time used for the simulation
`metadata`	JSON	Configuration details of the simulation, including noise rate `p`, number of `rounds`, error `sector`, chosen `decoder`, and its hyperparameters (`decoder_kwargs`)
Example `metadata` entry:

{
  "p": 0.001,
  "rounds": 3,
  "sector": "Z",
  "decoder": "bposd",
  "decoder_kwargs": [
    {
      "max_iter": 100,
      "ms_scaling_factor": 0.625,
      "osd_method": "OSD_CS",
      "osd_order": 15
    }
  ]
}

Source

Original paper: arXiv:2510.04521v1
Original data: Zenodo record 17220908
Blog Post: Iceberg-Quantum

Notes

All three CSVs published by the authors have been merged and converted to a unified Parquet format. The structure preserves the full fidelity of the original data while enabling efficient querying and analysis using libraries like Polars, PyArrow, or DuckDB.

Quick start

# Install dependencies
pip install polars aqora-cli pyarrow fsspec
import polars as pl
from aqora_cli.pyarrow import dataset
# Load dataset
df = pl.scan_pyarrow_dataset(
    dataset("aqora/fast-surgery-quantum-ldpc-codes", "v1.0.0")
)
df.head()